Study selection is summarised in Fig. 2. Of the final 80 papers, 22 studies modified attention patterns; 52, interpretational patterns; and 6, targeted both. All studies were used to address the range and effectiveness of interventions targetting attention and interpretational patterns. To address symptom reduction effects, only studies involving “clinical” participants (those reporting diagnoses of anxiety/depression) and “high-symptom” participants (those selected on the basis of symptoms above a clinical cut-off or in relation to other participants) were used. For these studies, we only included those that involved at least two intervention sessions as many early studies (even involving clinical/high-symptom participants) were designed to test for the plasticity of cognitive factors and causal links with symptom change in the short-term (within-session).
Interventions promoting helpful attention patterns
Attention Bias Modification interventions
Of 22 attention studies (8–28), nineteen (86%) used Attention Bias Modification (ABM) training to shift (and reinforce) maladaptive attention patterns away from threatening/negative to neutral or positive stimuli across multiple training trials. Common training tasks are the visual dot-probe and visual search training tasks (29) (Fig. 3). The visual dot-probe (and its’ variants) train the orienting of attention away from threatening/negative cues towards neutral/positive cues, by presenting response-probes more frequently behind neutral/positive stimuli.
The visual search task promotes goal-directed attention, by instructing participants to identify positive stimuli from an array of competing threatening/negative stimuli.
Across studies, ABM training was delivered between 1 and 13 sessions. Most studies employed a similar computerised task not designed to modify attention as an active comparison condition. For the visual dot-probe, this control condition involved responding to a probe that appeared with equal frequency behind a threatening/negative versus neutral/positive stimuli. In the visual search control task, participants searched for a 5-petalled flower from 7-petalled flowers. Most studies delivered ABM via computers, but smartphone methods were also trialled (30).
Ten ABM studies reported significantly greater reduction of attention biases for threatening/negative information in the intervention than the comparison condition (8, 12, 17, 18, 20–24, 26). Three other studies reported expected changes on attention bias but only when particular variants/versions of ABM were used (visual search(25), spatial cueing (23)); or under specific combinations of stimuli by exposure conditions (14).
Attention Flexibility interventions
Two studies(27, 28) (9%) aimed to improve general Attention Flexibility (AF, Fig. 3) using the Attention Training Technique (ATT) (31) and/or a Mindfulness-based intervention. ATT is designed to strengthen the ability to flexibly use and control attention through explicit instruction (31). Across training phases, participants engage in selective attention, attention switching and dividing their attention between neutral (e.g. sounds) stimuli in the environment. Mindfulness-based interventions can also target general attention regulation (32). Exercises may benefit concentration (the sustained aspect of attention), effortful attention-inhibition of distracting information, goal-directed attention control, and flexible switching of attention. While the focus is usually on neutral stimuli or internal sensations, these exercises also involve ‘sitting with’ more unpleasant sensations (e.g. pain) in non-judgmental way.
Only one of the two studies targeting AF collected measures to assess changes in attention patterns (28).
Participants receiving ATT or mindfulness showed similarly large (significant) pre-to-post increases in questionnaire measures of attention flexibility.
Symptom reduction effects of promoting helpful attention patterns
ABM interventions
Three multisession visual dot-probe ABM studies were conducted in clinical samples (Table 1). One involving young people with Social Anxiety Disorder (10) found small within-(intervention)-group reductions in symptoms from pre-to-post-intervention, but only ABM participants showed continued decreases to a 3-month follow-up. The absence of significant reductions in attention biases suggests that symptom changes were not driven by measurable changes in attentional patterns. Interestingly, this study noted that within adolescents allocated to receive ABM, those with higher trait attention control (reported by parents) showed significantly lower social anxiety symptoms at post-intervention. ABM was also assessed as a way of augmenting the effects of Cognitive Behavioural Therapy (CBT) in young people with complex forms of anxiety (9). Those who received ABM showed large within-group symptom reduction with large differences to the control-training-plus-CBT group at post-intervention. The authors did not report scores on attention bias measures making it difficult to attribute large symptom reduction effects to changes in attention patterns. Working with young people with depression (8), one study reported large within-group symptom reduction in the intervention group and a large group difference with control participants post-intervention (effects that persisted to a 12-month follow-up). Importantly, there were greater reductions in attention bias scores among those receiving ABM than the control condition, suggesting that symptom change could be due to attention change.
Table 1
Promoting helpful attention patterns in participants with clinical symptoms (in bold) and high symptom scores (not bold). Where studies do not give separate demographic information for each group, these are combined in a single cell. Where effect sizes were not reported or could not be calculated, these are labelled as Not Reported (NR). Green highlighted rows reflect studies showing large within group symptom reduction and at least medium sized between group effects. Orange highlighted rows reflect studies showing near large within group symptom reduction and small sized or non-reported between group effects.
Authors (year)
|
Presenting problem of sample
|
Intervention group number of participants; gender proportion, ethnicity, age range
|
Control group number of participants; gender proportion, ethnicity, age range
|
Number of sessions
|
Nature of control group
|
Within group effect size of pre-to-post primary symptom measure change
|
Between-group effect size of post-intervention primary symptom measure
|
Attention Bias Modification Training
|
Ollendick et al., 2019
|
Social Anxiety Disorder
|
N = 29;
31% male;
79% Caucasian;
Mean age = 14.34
|
N = 29;
28% male;
90% Caucasian;
Mean age = 14.24
|
10
|
Control training
|
0.25
|
0.09
|
Rieman et al., 2013
|
Anxiety disorders
|
N = 21;
48% male
100% Caucasian;
Mean age = 15.43
|
N = 21;
48% male;
90% Caucasian
Mean age = 15.71
|
NR but > 1
|
Control training
|
1.43
|
0.64
|
Yang et al., 2016
|
Major Depressive Disorder
|
N = 23;
48% male;
Ethnicity: NR
Mean age = 15.09
|
N = 22;
41% male;
Ethnicity: NR
Mean age = 14.82
|
12
|
Control training
|
2.77
|
0.63
|
Maoz et al., (2013)
|
Social Anxiety symptoms
|
N = 24
Males: 17%
Ethnicity: N/R
Mean Age = 22.96
|
N = 26
Males: 23%
Ethnicity: N/R
Mean Age = 22.42
|
4
|
Control training
|
0.07
|
0.05
|
Yao et al., (2015)
|
Social Anxiety symptoms
|
N = 23
Males: 43.5%
Ethnicity: 100% Chinese
Mean Age = 20.87
|
N = 23
Males: 26.1%
Ethnicity: 100% Chinese
Mean Age = 20.09
|
4
|
Control training
|
0.03
|
0.04
|
Pan et al., 2019
|
Social Anxiety symptoms
|
N = 20;
10% male;
100% Chinese:
Mean age = 20.05
|
N = 20;
15% male;
100% Chinese;
Mean age = 19.94
|
8
|
Control training
|
0.49
|
0.31
|
Liang (2016)
|
Social Anxiety symptoms
|
N(100ms group) = 18
Males: 27.8%
Ethnicity: 100% Taiwanese
Mean Age = 18.6
N(500ms group) = 18
Males: 33.3%
Ethnicity: 100% Taiwanese
Mean Age = 19.7
|
N(control100ms) = 18
Males: 22.2%
Ethnicity: 100% Taiwanese
Mean Age = 18.8
N(control500ms) = 18
Males: 27.8%
Ethnicity: 100% Taiwanese
Mean Age = 18.9
|
8
|
Control training at 100 and 500ms
|
1.23 (100ms)
0.89 (500ms)
|
1.10
(100ms)
0.71 (500ms)
|
Fitzgerald et al., (2016)
|
Social Anxiety symptoms
|
N = 61
Males: 23%
96.6% Caucasian
Mean Age = N/R
|
N = 59
Males: 27%
96.6% Caucasian
Mean Age = N/R
|
4
|
Control training
|
0.05
|
0.18
|
Bowler et al., (2017)
|
Anxiety symptoms
|
N = 22
Males: 32%
Ethnicity: N/R
Mean Age = 19.14
|
N = 24
Males: 31.8%
Ethnicity: N/R
Mean Age = 18.83
|
8
|
Control training
|
0.30
|
0.16
|
De Voogd et al., (2017)
|
Anxiety and depressive symptoms
|
N = 38
Males: 36.8%
Ethnicity: Dutch
Mean Age = 14.73
|
N = 32
Males: 34.4
Ethnicity: Dutch
Mean Age = 14.31
|
8
|
Control training
|
0.26
|
0.24
|
Baert et al., 2010
|
Depressive symptoms
|
N = 25
Males: 8%
Ethnicity: N/R
Mean Age = 19.88
|
N = 23
Males: 8.6%
Ethnicity: N/R
Mean Age = 20.09
|
12
|
Control training
|
-0.10
|
-0.73
|
Mastikhina
(2017)
|
Depressive symptoms
|
N = 24
Males: 13.3%
Ethnicity: Caucasian: 46.7%
Mean Age = 20
|
N = 21
Males: 13.3%
Ethnicity: Caucasian: 46.7%
Mean Age = 20
|
4
|
Control training
|
0.50
|
0.33
|
Wells (2009)
|
Depressive symptoms
|
N = 13
Males: N/R
Ethnicity: N/R
Mean Age = 19.1
|
N = 11
Males: N/R
Ethnicity: N/R
Mean Age = 19.1
|
4
|
Control training
|
0.52
|
NR
|
Yang et al., (2015)
|
Depressive symptoms
|
N = 27
Males: 25.9%
Ethnicity: N/R
Mean Age = 19.44
|
N = 27
Males: 37%
Ethnicity: N/R
Mean Age = 19.52
|
8
|
Control training
|
1.50
|
1.38
|
Attention flexibility interventions
|
Haukaas et al., (2018)*
|
Anxiety and depressive symptoms
|
N = 40
Males: 25%
Ethnicity: N/R
Mean Age = 22.7
|
N = 41
Males: 25%
Ethnicity: N/R
Mean Age = 22.7
|
3
|
Mindful Self-Compassion
|
Anxiety:
0.71
Depression: 0.53
|
0.36
|
Haukaas et al., (2018)*
|
Anxiety and depressive symptoms
|
N = 41
Males: 25%
Ethnicity: N/R
Mean Age = 22.7
|
N = 40
Males: 25%
Ethnicity: N/R
Mean Age = 22.7
|
3
|
Attention training techniques
|
Anxiety:
0.54
Depression: 0.308
|
0.36
|
Shomaker et al., (2017)
|
Depressive symptoms
|
N = 17
Males: 0%
70.6% Caucasian; Mean Age = 15.01
|
|
6
|
Group based cognitive therapy
|
1.96
|
0.56
|
* Study compared attention training technique against a mindfulness-based intervention; each of these groups is entered as a different row to show within-group changes |
Eleven ABM studies were conducted in high-symptom participants (Table 1). Seven involved young people with social or general anxiety symptoms (11–14, 16–18). With one exception(14), symptom reduction effects from pre-to-post intervention in the ABM training condition were small across studies (Cohen’s d = 0.03–0.49). Two (20, 21) (of the 4) studies targeting depression reported moderate-to-large size reduction in symptoms in the ABM condition (at post-intervention and follow-up time-points) that were linked with significant changes in attention bias. However only one of these reported a significant difference to their comparison condition at post-training and follow-up assessments (20). One study reported no differential reduction of depression symptoms in the ABM compared to control group (15), and unexpectedly, one study (19) found greater symptom decreases in control participants.
AF interventions
In a comparison of ATT with mindfulness in young people with high anxiety/depression symptoms (28), both groups showed medium-sized improvements, which were maintained at 6 months. Improvement in questionnaire reports of attention flexibility, significantly predicted treatment response. The second study(27) compared a mindfulness-based intervention with CBT in young people with depressive symptoms at-risk for Type 2 diabetes. Greater symptom reduction occurred in the mindfulness group than the CBT group at post-treatment (Cohen’s d = 0.56) and at six months follow-up (Cohen’s d = 0.69). No attention process measures were collected.
Interventions promoting helpful interpretation patterns
Cognitive Bias Modification of Interpretations interventions
Fifty-two studies (17, 18, 27, 30, 33–81) modified interpretations. Twenty-three (44%) included Cognitive Bias Modification of Interpretations (CBM-I) training. Similar to ABM, this uses repeated reinforcement learning to encourage the endorsement of positive (or benign) interpretations of ambiguous information, over-riding the tendency to infer threatening/negative explanations.
A common training task presents individuals with incomplete written ambiguous situations; completion of a word fragment resolves ambiguity in a positive/benign direction (Fig. 2). Most studies present incomplete emotionally-neutral situations as an active comparison condition. The number of training sessions across clinical, high-symptom and unselected studies ranged from single-sessions to 15.
Of 23 studies, 18 (78%) reported clear training effects in the CBM-I intervention condition, that is, either decreased negative interpretations or increased benign/positive interpretations from pre-to-post intervention or relative to the comparison condition at post-intervention. However, most studies used a measure of interpretational style that was structurally similar to the training task and could reflect demand effects. Where studies assessed transfer effects using a different measure of interpretation style, training effects in the CBM-I condition were mixed (33, 37, 54).
Cognitive Restructuring interventions
Thirty-one studies (59%) used Cognitive Restructuring (CR). This uses explicit instruction (Fig. 2) to encourage individuals to generate alternative explanations for situations and to consider evidence for and against each explanation. CR is a routine component in many CBT protocols for youth anxiety/depression, but can be used as a standalone intervention. Varying between 8–16 sessions, CR can be administered in individual or group sessions, face-to-face or remotely. One study (74) used Cognitive Reappraisal training, a version of CR focused on teaching participants to re-interpret distressing situations, sometimes through a third-party perspective (“psychological distancing”).
Seven of the 31 studies collected measures to inform changes in cognitive patterns, but none directly measured interpretation style. Yet, all 7 showed expected changes in the intervention compared to the comparison condition, which could reflect the products of increased positive/benign interpretation of daily situations. One reported changes in adolescents’ estimation of certain anxiety-provoking events (40). Another reported reductions in irrational beliefs(61). Decreases in automatic negative thoughts and increases in automatic positive thoughts (52) as well as decreases in self-negative statements (46, 62), and increases in positive cognitions around hypothetical stressful situations (79) were reported.
Symptom reduction effects of promoting helpful interpretation patterns
CBM-I interventions
Two studies (Table 2) delivered multisession CBM-I training to young people with clinical depression (33, 34), reporting small-to-moderate symptom reduction changes in the intervention group (Cohen’s d = 0.02/0.51). There were also small differences post-intervention with the control condition (Cohen’s d = 0.10/0.32).
Table 2
Promoting helpful interpretation patterns in participants with clinical symptoms (in bold) and high symptom scores (not bold). Where studies do not give separate demographic information for each group, these are combined in a single cell. Where effect sizes were not reported or could not be calculated, these are labelled as Not Reported (NR). Green highlighted rows reflect studies showing large within group symptom reduction and at least medium sized between group effects. Orange highlighted rows reflect studies showing near large within group symptom reduction and small sized or non-reported between group effects.
Authors (year)
|
Presenting problem of sample
|
Intervention group number of participants; gender proportion, ethnicity, age range
|
Control group number of participants; gender proportion, ethnicity, age range
|
Number of sessions
|
Nature of control group
|
Within group effect size of pre-to-post symptom measure change
|
Between-group effect size of post-intervention symptom measure
|
Cognitive Bias Modification of Interpretations training
|
Micco & Henin, 2014
|
Major Depressive Disorder
|
N = 23;
30% male; Ethnicity: N/R
Mean Age = 17.7
|
N = 22;
23% male; Ethnicity: N/R
Mean Age = 13.2
|
4
|
Control training
|
0.51
|
0.10
|
LeMoult et al., 2018
|
Major Depressive Disorder
|
N = 24;
21% male;
67% Caucasian:
13–17 years
|
N = 22;
99% male;
59% Caucasian:
13–17 years
|
6
|
Control training
|
0.02
|
0.32
|
Bowler et al., 2017
|
Anxiety symptoms
|
N = 26;
42% male;
Ethnicity: NR
Mean age = 19.0
|
N = 24;
32% male;
Ethnicity: NR
Mean age = 18.8
|
8
|
Control(ABM) training
|
0.59
|
0.22
|
Khalili-Torghabeh et al., 2014
|
Social anxiety symptoms
|
N = 18;
30% male;
100% Iranian; Mean age = 23.3
|
N = 17;
35% male;
100% Iranian; Mean age = 22.4
|
4
|
Control training
|
1.56
|
1.15
|
Butler et al., 2015
|
Social anxiety symptoms
|
N = 20
30% Male;
85% Caucasian
Mean age = 20.9
|
N = 20
30% Male;
95% Caucasian
Mean age = 24.4
|
3 sessions (in addition to CBT)
|
Control training (in addition to CBT)
|
0.79
|
0.20
|
Klein et al., 2018
|
Social anxiety symptoms (in mild intellectual disability)
|
N = 33;
36% male; Ethnicity: N/R
Mean Age = 14.4
|
N = 36;
31% male; Ethnicity: N/R
Mean Age = 13.2
|
5
|
Control training
|
NR for post-intervention
1.60 at 10-week follow up
|
0.10
|
Smith et al., 2019
|
Dysphoria
|
N = 36;
17% male;
67% Caucasian;
Mean age = 19.3
|
N = 36;
22% male;
69% Caucasian;
Mean age = 18.9
|
4
|
Psycho-education videos
|
0.78
|
0.12
|
DeVoogd et al., 2017
|
Anxiety and depression symptoms
|
N = 44;
41% male;
Ethnicity: N/R;
Mean age = 15.8
|
N = 30;
33% male;
Ethnicity: N/R; Mean age = 15.5
|
8
|
Control training
|
Anxiety: 0.22
Depression: 0.25
|
Anxiety: 0.18
Depression: -0.14
|
Cognitive Restructuring delivered as part of CBT
|
Waters & Donaldson, 2008
|
Generalised Anxiety Disorder
|
N = 4;
0% male;
100% Caucasian:
14–17 years
|
N/A
|
10
|
No control group
|
NR
|
N/A
|
Albano et al., 1995
|
Social Anxiety Disorder
|
N = 5;
60% male;
100% Caucasian:
13 to 17 years
|
N/A
|
16 (in a group)
|
No control group
|
NR
|
N/A
|
Taheri et al., 2018
|
Social Anxiety Disorder
|
N = 30
43% male;
Ethnicity: NR
Mean age = 21.99
|
10 (in a group)
|
Behavioural activation
|
4.28
|
0.17
|
Micco et al., 2007
|
Panic disorder
|
N = 12
17% male;
100% Caucasian:
13–17 years
|
N = 9
11% male;
100% Caucasian
13–17 years
|
11
|
Wait-list
|
0.98
|
NR
|
Ginsburg & Drake, 2002
|
Anxiety disorders
|
N = 4;
NR% male;
100% African-American;
14–17 years
|
N = 5; NR% male;
100% African-American;
14–17 years
|
10 (in a group)
|
Group attention support
|
0.28
|
0.26
|
Topooco et al., 2018
|
Major Depressive Disorder
|
N = 33
6% male;
Ethnicity: N/R
Mean Age = 17.2
|
N = 37
5% male;
Ethnicity: N/R
Mean Age = 16.9
|
8 (Internet)
|
Attention support
|
1.58
|
0.71
|
Rosenberg et al., 2011
|
PTSD with concurrent depression symptoms
|
N = 12;
25% male;
83% Caucasian;
Mean Age = 16
|
N/A
|
12–16
|
No control group
|
1.43
|
N/A
|
Sportel et al., 2013
|
Social anxiety symptoms
|
N = 84
33% male;
Ethnicity: NR
Mean age = 14.1
|
N = 70
23% male;
Ethnicity: NR
Mean age = 14.1
|
10
|
No intervention
|
0.17 (post-test)
0.85 (6 months)
|
0.16 (post-test)
0.41 (6 months)
|
Wessel & Mersch, 1994
|
Test anxiety
|
Intervention N = 22;
Control N = 8
46% male;
Ethnicity: NR
Mean age = 16.5
|
10 (in a group)
|
No intervention
|
1.00
|
0.80
|
Hains et al., 2001
|
Anxiety symptoms in young people with Type 1 Diabetes
|
N = 6
50% male;
83% European-American
Mean age = 14.16
|
No control group
|
8
|
N/A
|
NR
|
NR
|
Clarke et al., 2001
|
Depression symptoms in offspring of depressed parents
|
N = 40
47% male;
82% Caucasian;
Mean age = 14.4
|
N = 47
35% male;
96% Caucasian
Mean age = 14.7
|
15 (in a group)
|
Treatment as usual
|
0.85
|
0.48
|
Deady et al., 2016
|
Depression symptoms and co-occurring alcohol problems
|
N = 60;
40% male;
Ethnicity: NR
Mean age = 21.9
|
N = 60;
41% male;
Ethnicity: NR
Mean age = 21.6
|
4
|
Attention and support by therapist
|
1.09
|
0.71
|
Wright et al., 2017
|
Depression symptoms
|
N = 41;
27% male;
100% Caucasian;
Mean age = 15.5
|
N = 42;
41% male;
98% Caucasian;
Mean age = 15.2
|
8 (Internet)
|
Attention and support by therapist
|
0.45 (4 months)
0.39 (12 months)
|
0.25 (4 months)
0.15 (12 months)
|
Topooco et al., 2019
|
Depression symptoms
|
N = 35
9% male
Ethnicity: N/R
Mean age = 17.5
|
N = 35
0% male
Ethnicity: N/R
Mean age = 17.5
|
8 (Internet)
|
Attention and support by therapist
|
1.46
|
0.86
|
Shomaker et al., 2017
|
Depression symptoms in young people at-risk for Diabetes
|
N = 16
0% male;
69% Caucasian:
Mean age = 15.0
|
N = 17
0% male;
71% Caucasian:
Mean age = 15.0
|
6 (in a group)
|
Mindfulness-based intervention
|
0.76
|
-0.25
|
Cognitive Restructuring (standalone)
|
Fernandez et al, 2017
|
Anxiety symptoms
|
No details reported
|
No details reported
|
5 (in a group)
|
No intervention
|
1.88
|
0.87
|
Sweeney & Horan, 1982
|
Performance anxiety
|
N = 9; no other details reported
|
N = 9; no other details reported
|
6 (in a group)
|
Wait list control
|
0.77
|
0.29
|
Fremouw, 1978
|
Speech anxiety
|
N = 12;
%male: NR
Ethnicity: NR
18–24 years
|
N = 11;
%male: NR
Ethnicity: NR
18–24 years
|
5 (in a group)
|
No intervention
|
0.72
|
0.22
|
Gross et al., 1982
|
Speech anxiety
|
Overall N = 63;
27% male
Ethnicity: NR
18–22 years
|
4 (in a group)
|
No intervention
|
2.43
|
2.12
|
Bistline et al., 1980
|
Test anxiety
|
N = 9;
%male: NR
Ethnicity: NR
18–22 years
|
N = 11;
%male: NR
Ethnicity: NR
18–22 years
|
5 (in a group)
|
No intervention
|
0.84
|
0.83
|
Decker et al, 1981
|
Test anxiety
|
N = 9
%male: NR
Ethnicity: NR
Mean age = 19
|
N = 8
%male: NR
Ethnicity: NR
Mean age = NR
|
4 (in a group)
|
No intervention
|
NR
|
NR
|
Wise & Haynes, 1983
|
Test anxiety
|
N = 14
%male: NR
Ethnicity: NR
18–24 years
|
N = 13
%male: NR
Ethnicity: NR
18–24 years
|
5 (in a group)
|
No intervention
|
NR
|
NR
|
Arnkoff, 1986
|
Test anxiety
|
N = 17;
%male: NR
Ethnicity: NR
18–24 years
|
N = 16;
%male: NR
Ethnicity: NR
18–24 years
|
4 (in a group)
|
No intervention
|
NR
|
NR
|
Clore & Scott, 2006
|
Depression symptoms
|
N = 10;
20% male;
90% European-American
Mean age = 21.7
|
N = 10;
20% male;
90% European-American
Mean age = 20.5
|
3
|
Increasing positive self-thoughts
|
1.69
|
0.25
|
Six studies applied CBM-I to young people with high symptom scores (Table 2). Only one reported large symptom improvement in the CBM-I group, and equally large differences compared to a control condition, with expected changes in interpretational style (35). Three studies of general anxiety(18), social anxiety (36) and dysphoria (38) showed medium-sized symptom reduction in the CBM-I group (Cohen’s d = 0.59–0.79) and small between-group differences with various comparison conditions post-intervention (Cohen’s d = 0.12–0.22). Finally, two studies targeting social anxiety(37) or anxiety/depression (17) reported small within-group symptom reduction and small post-intervention differences with the comparison condition. However, for one, CBM-I training effects on symptoms were more apparent at 10-weeks (Cohen’s d = 1.60)(37), with expected changes in positive interpretation.
CR interventions
Seven studies assessed CR techniques within CBT in young people with clinical anxiety and/or depression, as the primary condition or co-morbid with PTSD (Table 2). Three were case series (39, 43, 81), where no data on symptom measures was reported, or were based on fewer than 5 participants. Two studies targeting anxiety disorders (social anxiety (41), panic (40)) showed large reductions in symptoms from pre-to-post intervention in the CR condition. Micco and colleagues (40) recorded session-to-session change on anxiety, and noted a therapeutic gain following the first session of CR. However, in both studies, there was either no data reported from the (wait-list) control group, or the control group was another active intervention (behavioural activation (41)), in which case the between-group difference at post-intervention on anxiety symptoms was small. Using CR to target depression, two studies revealed large within-group symptom reduction effects (42, 44), but only one employed a comparison condition (comprising monitoring and non-specific counselling), and reported a moderate sized post-intervention difference in depressive symptoms (42).
Nine studies employed CR (within CBT) in young people with high symptom levels. Two assessed CR effects on anxiety symptoms with one finding weak (45), and the other strong (46), within-group reduction effects. The study reporting the stronger within-group changes noted a reduction in negative self-statements (46). A case series aiming to reduce anxiety symptoms in young people with a chronic health condition (47) did not report whole-sample mean changes but all 6 participants improved across treatment. For the 6 studies assessing depression symptoms (with one targeting depression in a sample at-risk for a chronic health condition (27)), 4 reported large within-intervention-group pre-to-post symptom reduction effects and medium-to-large between-group differences with a comparison condition (treatment as usual or attention support) at post-intervention (42, 48–50). The two other studies reported weak or medium-sized symptom reduction effects (27, 51); the one reporting smaller-sized changes assessed this at 4 months (51) so improvements may have become weak with time.
Nine studies examined standalone CR interventions in young people with high general anxiety/depression symptoms or with specific test, speech or performance anxiety (Table 2). Three noted significant reductions in anxiety measures from pre-to-post-intervention in the CR condition but did not report enough data to calculate effect sizes (58, 60, 61). Where effect sizes were reported, within-group symptom reduction were medium to large (Cohen’s d = 0.72–2.43). Two studies that reported large effect sizes also found expected changes in automatic negative thoughts and negative/positive self-statements in the CR group (52, 62). Comparisons with waitlist/no-intervention conditions across studies showed weak to large between-group differences at post-intervention. One study noted that individuals with lower purposeful engagement (the reduced tendency to attend/engage with unpleasant thoughts) benefited more from CR techniques than those receiving the comparison condition (44).
Amplifying the effects of attention and interpretation interventions
Combined interventions
Six studies jointly targeted attention and interpretation patterns in reducing anxiety/depression. Four involved clinical participants. The first (82) delivered a web-based intervention combining CR and ATT in reducing social anxiety. Effect sizes (Cohen’s d) for the intervention group were 0.72–0.82 on symptoms from baseline to a 4-month follow-up, and were significantly greater than changes reported in the wait-list control group. Piet and colleagues (83) combined mindfulness-based cognitive therapy with CBT in socially-anxious young adults, thus targeting AF and CR. Combining interventions yielded greater within-group symptom reduction but this increase was marginal compared to receiving one treatment (Cohen’s d = 0.20–0.33). O’Toole and colleagues (84) applied Emotion Regulation Therapy to young people with generalised anxiety disorder, cultivating AF (shifting and sustaining attention on a difficult experience) and CR abilities. Within-intervention-group reductions emerged on anxiety symptoms (Cohen’s d = 1.2–1.4), preceded and mediated by changes in cognitive reappraisal and decentring, a cognitive skill, inversely correlated with negative self-referential processing. Finally, a one-day group-based CBT package was delivered to 24 young people with clinical anxiety/depression (85). Of 8 workshop topics, one corresponded to CR and one to AF (within mindfulness). The study aimed to gather qualitative feedback from young people. One theme that emerged was that the wide selection of techniques enabled young people to learn a suitable technique.
Two case series (86, 87) piloted a combined ABM and CBM-I intervention to reduce high anxiety/depression symptoms. Neither included a comparison condition and neither was powered to detect significant within-group changes. The first showed reduction of social anxiety symptoms in around 80% of participants. Using a similar intervention but adapted/translated for young people with a history of victimisation in the UK and Nepal, social anxiety scores reduced only in UK participants (Cohen’s d = 0.81). Qualitative feedback across both studies suggested poor engagement with ABM than CBM-I training.
Additional strategies suggested by young people with lived experiences
Ten young people (15–24 years; 7 females; 6 White British) with varying severity of past anxiety and depression (and treatments) were consulted about managing unhelpful thinking styles in daily life. Several messages emerged.
All participants agreed that adopting alternative perspectives, such as those used in CR or CBM-I, was useful in daily life, especially using a third-person perspective.
“Yeah so you don’t think all the focus is on you…you could say something from a positive angle like they could be talking about somebody else not talking about you.” 21-year old female, Asian British.
“The therapy helped me by not having the negative thoughts.. making me aware of it… talking about this, is it really the person or is it me, or getting another perspective… rather than cutting the relationship has helped” 24-year old female, White European.
“Challenging those thoughts do[es] help, especially trying to find pieces of evidence to go against what I think… It is also helpful to get someone else to suggest alternative perspectives” 21-year old female, Chinese.
However, some noted obstacles in current interventions, underscoring a need to help young people discover methods for learning and implementing helpful cognitive patterns:
“In counselling, they try to think about what is and isn’t irrational. And try to think about all the positive explanations before you jump to negatives… if I can recognise it is happening this can be easier, but very often I don’t so it doesn’t help massively”. 17-year old female, White British.
One young person suggested using both attention and interpretation patterns to manage negative emotions:
"I think someone once told me that a thought only lasts for 8 seconds unless you chose to prolong it yourself. For me, I find it easier to engage it a little more and find out where it has come from, why do I think this, where has it come from, and then challenge the idea in my head” 19-year old male, White British.
Young people also described that the deliberate recall of positive past experiences or positive aspects of oneself could help to counteract negative thinking:
“May be if I try to point out the things that are good and maybe compare it to another piece of artwork I’ve done before. Or maybe think “well you’ve struggled on this but look now, you’ve managed to do this better than you’ve done before”” 15-year old female, White British.
“I write down things I appreciate about myself… Say if I cooked dinner for my friends, they didn’t give that much positive feedback, I might automatically assume my cooking wasn’t that good or I had not cooked enough food. But then if I wrote down that I actually cooked for them, I might feel better.” 21-year old female, Chinese.